Delirium: the quiet one is the worry

Clinical Rounds, the GPAtlas podcast. Mental Health (MH). 20 minutes. Published 31 August 2026. Free to listen.

Clinical source: our own Ocean article "Delirium", last reviewed 5 June 2026.

Transcript

Sarah: Hello, and welcome to GPAtlas Clinical Rounds. I'm Sarah, and this is the podcast where we take one thing you'll genuinely see this week and work out what to do with it.

Ben: And I'm Ben. Today, delirium. And specifically the version of it that doesn't shout.

Sarah: So. Mrs Cole is eighty-four, she lives alone, and the request is a home visit because she's not herself. When I get there she's in her chair, she's quiet, she answers me, she's polite. Her daughter says she's been sleepy for a couple of days and off her food. And the honest truth is that everything in that room is telling me she's fine, and there's a version of me on a busy Thursday that leaves.

Ben: Three things today. Why the quiet patient is the dangerous one. The two questions that make the diagnosis, which take about a minute between them. And then what to actually do, including the treatment that isn't a drug.

Sarah: So let's start with what's counter-intuitive, because it's the whole episode. When people picture delirium they picture the agitated patient. Pulling at things, shouting, trying to leave. And that version gets diagnosed, because it's impossible to ignore.

Ben: But our article is unambiguous about the subtype that matters. Hypoactive delirium, so drowsiness, withdrawal, slowed responses, is the commonest and most easily overlooked, with the worst outcomes. Commonest. Worst outcomes. And the line that should stay with you: the quiet patient who is no trouble is the one most often missed.

Sarah: And it's missed because it looks like other things we're comfortable with. Our article says hypoactive delirium is frequently mislabelled as depression, fatigue, or simply getting worse with age. All three of those are diagnoses that let everybody go home. And the third one is the worst, because getting worse with age isn't a diagnosis at all, it's a shrug.

Ben: The quiet one is the commonest. And it has the worst outcomes.

Sarah: So how do you catch it, in a home visit, in ten minutes, with a daughter watching you. Two questions. The first one isn't for the patient, it's for whoever brought you there.

Ben: It's called the Single Question in Delirium, the SQiD, and it is exactly what it sounds like. Is this person more confused than usual? That's it. And our article's framing is the useful part: a change from their baseline is the key diagnostic clue. So you're not assessing whether she's confused. You're assessing whether she's changed.

Sarah: And the second question is for the patient, and it's the one I have to make myself do, because it feels rude when someone is being perfectly pleasant. You test attention. Months of the year backwards. Our article says to actively test attention in any drowsy or withdrawn older patient rather than assuming they are settled. Assuming they're settled is exactly what a polite, quiet patient invites you to do.

Ben: And if you want the formal tool, it's the 4AT, which is the NICE-recommended bedside instrument: alertness, an abbreviated mental test, attention through months backwards, and acute change or fluctuation. The SQiD asked of the family adds the collateral. Between them that's a minute and a half of your consultation.

Sarah: Is she different from usual? Then: months of the year, backwards.

Ben: Now the distinction everybody worries about, delirium versus dementia. And the discriminator is attention. In delirium it's markedly impaired, that's the core deficit. In dementia it's usually preserved until late. Delirium is acute, it fluctuates, alertness is altered, and it's usually reversible.

Sarah: And there's one exception that will catch you, which our article names: Lewy body dementia, which also fluctuates. So the fluctuation you're relying on isn't unique. And the practical instruction it gives is the one I'd want a registrar to remember, because it stops you being paralysed by the diagnostic question. Where the diagnoses are hard to separate, manage the delirium first.

Ben: And the most dangerous version of this is the patient who already has dementia. Our article says a sudden deterioration in a person with dementia is delirium until proven otherwise, and never attribute an acute change simply to the dementia getting worse. Because dementia does not deteriorate over forty-eight hours. If it looks like it has, something else has happened.

Sarah: Dementia does not change in two days. Something else did.

Sarah: Right. So I've decided Mrs Cole is delirious. Now I have to find out why, and the answer is almost never one thing. Our article says delirium is usually multifactorial, and gives the mnemonic, PINCH ME. Pain. Infection. Nutrition. Constipation. Hydration. Medication and metabolic. Environment.

Ben: And two of those deserve singling out because they're where primary care actually wins. Constipation, which nobody examines for and which is enormously common in exactly this group. And medication, where our article says reviewing and deprescribing anticholinergics is one of the highest-value actions in primary care. Not a tidy-up. One of the highest- value things you can do.

Sarah: And the reason it's high value is the arithmetic of it. Each individual drug looks defensible. The amitriptyline for her sleep, the oxybutynin for her bladder, the sedating antihistamine she buys herself. Our article's phrase is that when several are co- prescribed the combined burden rises sharply. Nobody prescribed a delirium. Four people each prescribed a reasonable thing.

Ben: With one caution attached: stop or reduce the anticholinergics, sedatives, opioids and other culprits where it's safe, but avoid abrupt withdrawal. Stopping a long-standing benzodiazepine overnight in an eighty-four year old creates a second problem on top of the first.

Sarah: Nobody prescribed a delirium. Four people prescribed a reasonable thing.

Ben: And then the test that we get wrong more than any other in this scenario. Our article is direct: asymptomatic bacteriuria is so common in older people that by age eighty around half of those in care carry bacteria in the urine without infection, so a positive urine dipstick does not diagnose a urinary tract infection. And it says the UK Health Security Agency advises against dipsticks in the over-sixty-fives.

Sarah: And that's a hard one, socially, because the family often wants the dipstick. It's quick, it's visible, it produces an answer, and it leads to antibiotics, which feels like doing something. Whereas what our article is telling you is that a positive stick in that patient tells you almost nothing, and the risk is that you stop looking. You've got your answer, you've given trimethoprim, and the constipation and the amitriptyline are still there.

Ben: There is one bedside test that should happen every time, though. Glucose. Our article calls it an immediate bedside test in every confused patient, to exclude hypoglycaemia and marked hyperglycaemia. That one's non-negotiable and it takes thirty seconds.

Sarah: Do not dipstick her. Do check her glucose.

Sarah: So, management. And the headline is that the first-line treatment is not a drug. For distress or risk, our article says use verbal and non-verbal de-escalation first, and it adds this is first-line, not medication. So that's orientation, familiar people, light, glasses on, hearing aids in, reducing the noise, and treating the pain and the constipation you found.

Ben: And there's a sentence about sedation that ought to be on the wall of every out-of-hours car. Sedation does not treat delirium. Antipsychotics are a short-term safety measure for severe distress or risk only, and they can prolong the episode and increase harm. So the drug isn't a treatment that's been escalated to. It's a containment measure with a cost.

Sarah: If you do need it, the numbers are small and worth knowing. Haloperidol, typical starting dose half a milligram, oral or intramuscular, lowest effective dose for the shortest time, per your local protocol. And avoided in Parkinson's disease, in Lewy body dementia, and where the QT is prolonged.

Ben: And where antipsychotics are contraindicated, so in Parkinson's or Lewy body dementia, lorazepam half a milligram, oral or intramuscular, is the preferred option, with the honest caveat our article attaches: benzodiazepines can themselves worsen or prolong delirium, so use them cautiously. There isn't a clean answer in that corner, and pretending otherwise is how people get harmed.

Sarah: Sedation is not treatment. It is containment, and it has a cost.

Ben: Two clear admission triggers, because managing delirium at home is reasonable and sometimes it isn't. If the cause is unclear after your initial primary-care assessment, admit for investigation, because delirium is a medical emergency. And if it's unsafe to manage at home, so inadequate supervision, dehydration, falls risk, arrange admission.

Sarah: And the legal side, which comes up the moment you're deciding anything without her full agreement. Under the Mental Capacity Act, a delirious patient often lacks capacity for specific decisions. You assess and document capacity for each decision, act in their best interests, and involve family and any advance statement or lasting power of attorney. Per decision, not once for the person.

Ben: Cause unclear after your assessment means admit. It is an emergency.

Sarah: And then the part that gets forgotten, because by the time she's better everybody has moved on. Most delirium settles within days. But our article says around one in five persists beyond a week, and full recovery can take weeks to months. So the family who were told she'd be back to normal in a few days need a more honest timeline than that.

Ben: And two follow-up actions. Document the diagnosis in the GP record, which sounds administrative but is the thing that makes the next episode recognisable. And after recovery, reassess her cognition, because an episode of delirium is associated with a higher future risk of dementia, so consider a memory-clinic referral if cognitive concerns persist.

Sarah: Which reframes the whole visit, really. It isn't just an acute problem you fix and close. It's a marker. She's told you something about her brain's reserve that you didn't know this morning, and the record should say so.

Ben: One in five is not better in a week. Say that to the family.

Sarah: Let me go back to Mrs Cole, because I want to be honest about how that visit actually goes. I ask the daughter whether this is different, and she says yes, definitely, since about Tuesday. I ask Mrs Cole the months of the year backwards and she gets to September and stops and looks at me. And that's the diagnosis, in about ninety seconds, in a chair, with no equipment.

Sarah: Then I go through PINCH ME out loud, because I find it stops me settling on the first plausible thing. Pain, she's got her back. Infection, possible but I'm not going to prove it with a stick. Nutrition, she's not eating. Constipation, five days. Hydration, poor. Medication, and here's her repeat list with codeine on it. Environment, she's alone and it's been hot.

Sarah: And what strikes me every time I do that is how the answer stops being a diagnosis and starts being a list. There isn't one cause to find and fix. There are six things that are each a bit wrong, and my job is to make four of them less wrong by tomorrow. That's a different kind of medicine to the one we're trained to want.

Ben: And it's the kind primary care is uniquely placed to do, which is worth saying. You have her repeat list, you know her baseline, and you can go back tomorrow. A hospital has none of those three. That's why our article puts the medication review so high, and why the admission triggers are about safety and uncertainty rather than about needing a hospital to treat delirium as such.

Sarah: It stops being a diagnosis and becomes a list. Make four of them less wrong.

Ben: And it's worth ending on why this matters beyond the episode itself. Our article notes that around one in five persists beyond a week, that full recovery can take weeks to months, and that an episode of delirium is associated with a higher future risk of dementia.

Ben: So this isn't a self-limiting confusion that resolves and leaves nothing behind. It's a marker of vulnerability, and the thing that makes it visible later is whether anybody wrote it in the record at the time. Which is a two-second action with a multi-year payoff, and it's the one most likely to be skipped on a home visit at the end of a shift.

Sarah: And practically, the reassessment afterwards is ours too. Once she's better, you reassess her cognition, and if concerns persist you consider a memory clinic referral. That's a follow-up appointment that nobody will book unless we do.

Ben: Write it in the record. That is what makes the next one visible.

Ben: Exam corner. An eighty-four year old woman with mild dementia is brought in by her daughter, who says she's been much more confused for three days. She's drowsy and withdrawn. She's on codeine for back pain, and she hasn't opened her bowels for five days. A urine dipstick is positive for nitrites and leucocytes. What is the most appropriate initial action? A, prescribe trimethoprim for a urinary tract infection. B, attribute the change to progression of her dementia. C, assess for constipation and review her medication, treating the delirium as multifactorial. Or D, start haloperidol for agitation.

Ben: It's C. She has a five-day history of constipation and she's on codeine, and delirium is usually multifactorial, so PINCH ME points you at pain, constipation and medication all at once. A is the trap: a positive dipstick in an eighty-four year old does not diagnose infection, because asymptomatic bacteriuria is so common at that age, and treating it stops you looking. B is specifically warned against, since a sudden deterioration in dementia is delirium until proven otherwise. And D is wrong twice over: she's withdrawn rather than agitated, and sedation doesn't treat delirium anyway.

Sarah: So. Three things for Monday. One. On any drowsy or withdrawn older patient, ask the family whether this is a change, and test attention rather than assuming she's settled. Two. Don't dipstick the over-sixty-fives, and do check a glucose. And three. Look at the repeat list, because the anticholinergic burden is one of the highest-value things you can act on, and nobody deliberately caused it.

Sarah: One thing to reflect on, if you're logging this. Think of the last patient you saw who was described to you as just quiet, or a bit flat, or no trouble. Did anybody test their attention? The full transcript and the references are on the episode page. Ben and I are synthetic voices. The medicine isn't. See you next week.

More from Clinical Rounds

Every episode of Clinical Rounds, free to listen with a full transcript.